fix: recursively traverse parent chain for inherited MediaBox

[emerose]

Summary

PDF specification (sec 7.7.3.4, p 80) allows /MediaBox to be inherited from any ancestor in the page tree, not just the immediate parent. The previous fix in v4.7.1 only checked one level of parent, which failed for PDFs where /MediaBox is defined at the grandparent level or higher (common in multi-level page trees).

Problem

Certain valid PDFs fail with:

RuntimeError: could not find the page-dimensions: {"/Type": "/Page"}

These PDFs have /MediaBox defined at the root of the page tree (e.g., grandparent), not on individual pages or their immediate parents.

Example PDF Structure (EU-lex document)

Page (xref 1)
  └─ Parent: Pages node (no /MediaBox)
       └─ Parent: Root Pages node (/MediaBox [0 0 595 842]) ← Here!

Solution

Recursively traverse the parent chain (up to 10 levels to prevent infinite loops in malformed PDFs) until /MediaBox is found or the root is reached.

for(int depth = 0; depth < 10 && current.hasKey("/Parent"); depth++)
{
    QPDFObjectHandle parent = current.getKey("/Parent");
    if(parent.hasKey("/MediaBox"))
    {
        // Extract MediaBox from this ancestor
        ...
    }
    current = parent;
}

Testing

• All existing tests pass (16/16)
• Added test case: deep-mediabox-inheritance.pdf (EU-lex document with MediaBox at grandparent level)
• Tested with multiple real-world PDFs that previously failed

Fixes

Fixes #175

Signed-off-by: Sam Quigley sq@emerose.com

[github-actions]

✅ DCO Check Passed

Thanks @emerose, all your commits are properly signed off. 🎉

[mergify]

Merge Protections

Your pull request matches the following merge protections and will not be merged until they are valid.

🟢 Enforce conventional commit

Wonderful, this rule succeeded.

Make sure that we follow https://www.conventionalcommits.org/en/v1.0.0/

• title ~= ^(fix|feat|docs|style|refactor|perf|test|build|ci|chore|revert)(?:\(.+\))?(!)?:

[PeterStaar-IBM]

@emerose Love it! Thanks for the code + the test-case.

You do have to add the ground-truth files too (they get created otherwise)

[dolfim-ibm]

lgtm

[PeterStaar-IBM]

wonderful!